\begin{table}[h!]
\centering
\caption{Average Treatment Effect: Interview Scheduled}
\centering
\begin{tabular}[t]{lccc}
\toprule
& \multicolumn{3}{c}{\textit{Sample}}\\
  & Pooled & Firm & OECD Bureaucrat\\
\midrule
Jake Miller & \num{-0.082}* & \num{-0.143} & \num{-0.044}\\
 & (\num{0.050}) & (\num{0.102}) & (\num{0.053})\\
\midrule
Num.Obs. & \num{173} & \num{57} & \num{116}\\
R2 & \num{0.016} & \num{0.032} & \num{0.006}\\
\bottomrule
\multicolumn{4}{l}{\rule{0pt}{1em}* p $<$ 0.1, ** p $<$ 0.05, *** p $<$ 0.01}\\
\end{tabular}
\begin{quote}
    \footnotesize \textit{Note:} This table shows the OLS regression results for the average treatment effect of outreach gender on whether the elite scheduled an interview. As pre-registered, the independent variable is the outreach gender -- 1 if ``Jake Miller'' and 0 if ``Mary Williams.'' The dependent variable is whether the elite scheduled and attended an interview, taking the value of 1 if yes, 0 if not. I test these results in the pooled sample of firm representatives and OECD bureaucrats (Model 1), amongst only the sample of firm representatives (Model 2), and amongst only the sample of OECD bureaucrats (Model 3).
\end{quote}
\label{a1}
\end{table}
